Unemployment Insurance as a Subsidy to Risky Firms

Transcription

1 Unemployment Insurance as a Subsidy to Risky Firms by Bernardus Van Doornik, Dimas Fazio, David Schoenherr & Janis Skrastins Discussion - June 2018 Jen Brown

2 Research question: Does unemployment insurance affect workers labor supply to risky firms? STEP BACK Given this question, what is the ideal experiment? Randomly assign (otherwise identical) workers to more/less generous UI programs Randomly assign those workers to (otherwise identical) safe/risky firms Less generous UI More generous UI Safe Firms firms a b Risky firms c d Then compare outcomes (employment, wages)

3 Research question: Does unemployment insurance affect workers labor supply to risky firms? SUMMARY The paper exploits an unexpected change in UI eligibility in Brazil in 2015 The empirical setting is attractive Key feature of the reform: only some workers were affected by the change in UI eligibility Identifying assumption: unaffected workers were not differentially influenced by other external changes, compared to affected (but otherwise similar) workers The data are also enviable worker-firm matched data covering formal wages, age, etc. of 50MM workers

4 Less generous UI reduces labor supply RESULTS The paper examines equilibrium outcomes to learn about labor supply: UI reform lead to relatively lower total formal employment and higher wages for affected workers RISK Classifying firms by risk, the negative employment and positive wage effects of UI reform are especially strong at risky firms The authors proxy for employment risk with: Historical layoff rates Credit spreads Default provisions with the Central Bank Falato & Liang 2015 Agrawal & Matsa 2013 Hotchkiss 1995 o There is some survey evidence that workers do perceive these differences across firms (Brown & Matsa 2016)

5 The paper exploits other exogenous shocks: this time, shocks to firm performance EXOGENOUS SHOCKS Generous UI might lead to riskier firm choices (Kim 2012; Agrawal & Matsa 2013) Authors address the potential endogeneity of firm risk using exogenous, severe weather-related shocks Risky firms have relationships with customers or suppliers affected by large weather events Shock unemployment risk through an aggregate demand channel

6 Wages at risky firm are higher for workers facing reduced UI eligibility HIGHER WAGE Compensating wage differential studies have often use data in which unobserved worker characteristic (e.g. productivity) may be correlated with unemployment risk Here, we see a cleaner setting, since the shock to the cost of unemployment is exogenous How does a back-of-the-envelope (or maybe better) calculation of lost UI payments compare to the increased wage?

7 Regardless of what the compensating differential is covering, the wage bump isn t sufficient HIGHER WAGE But, is the wage change only about employment risk? Van Doornik, Schoenherr & Skrastins (w.p. 2018) discuss formal and informal labors markets, and same UI reform If reform made the UI system harder to game, then higher wages are compensating for lost benefit of working informally while collecting UI, not unemployment risk per se Are riskier firms more likely to have been colluding with workers? How would this affect the historical layoff-related risk measure? Examine only firms/industries that cannot substitutes to informal labor?

8 How are the affected and unaffected workers different, beyond UI eligibility? TREATMENT VS CONTROL Table 1 provides some comparisons, but it would be helpful to see direct comparisons of workers age, wealth, education, occupation, industry, location by UI eligibility and firm type Do workers types appear in the same proportion inside risky vs. safe firms? After accounting for worker age, affected workers should have longer tenure? Are affected workers (little prior unemployment) more likely to be professionals with steep wage-tenure profiles? That is, are the unaffected workers (systematically) on a flatter section of a wage-tenure profile?

9 Heterogeneity in the effect of interest adds depth to the discussion HETEROGENEITY UI reform should have less bite when workers have savings Individual wealth may be hard/impossible to measure College grads may have greater liquid savings and UI benefits typically cover a smaller fraction of lost income (Chetty 2008) So, more educated workers employment and wage should be less sensitive to UI reform Age and/or tenure interactions with the UI reform might capture similar differences in sensitivity due to accumulated wealth

10 The performance of risky firms declines after UI reform FIRM VALUE Why? Do these firms now face a higher wage bill? Can this (almost mechanical) consequence be observed directly? Did these firms lose their high quality labor because an insufficient compensating wage differential? Is there a productivity measure that can proxy for worker quality? Do these firms undertake a real strategic change? Do they change their approach to value creation in the face of worker attrition or higher labor costs?